Modeling complex chemical reactions is a long-standing challenge in chemical reaction engineering. Group type analysis is commonly used for reducing the complexity of the reaction model while retaining sufficient chemical information for a particular application. This study evaluates a data-driven approach based on mathematical similarity as a new tool for identification of groups for use in group type reaction models. Data for dodecane cracking in the supercritical state over Zeolite Socony Mobil-5 (ZSM-5) was used as a test system. Mathematical similarity analysis of the raw data differentiated as many as five different groups. Adding synthetic data did not affect the groups identified by similarity analysis, indicating that the separation was not limited by the number of data points. Scaling or normalizing the data improved the separation of the similarity analysis. To test the data-identified groups, different types of reaction models were generated systematically, kinetic parameters regressed, and the resulting predictions compared with experimental data. The resulting reaction models consisted either of parallel or sequential reactions. As a general statement, the reaction models consisting of parallel reactions were more accurate than those consisting of sequential reactions. Additional tests showed that user defined groups could be added to those identified mathematically to improve the accuracy of predictions for target species without sacrificing overall accuracy. The similarity approach was applied to a data set consisting of catalytic dodecane cracking in the presence of water added to the reaction mixture (50 wt %). The proposed similarity analysis identified different groups in the presence and absence of water, indicating that the data-driven approach can be used to identify qualitative differences in the reaction pathways. It is therefore demonstrated that data-driven identification of group types represents a useful new tool for development of group type models.

The molecular structure of hydrochars produced from 13C-enriched glucose under various conditions has been elucidated based on advanced one- and two-dimensional (2D) 1H-13C and 13C-13C solid-state nuclear magnetic resonance (NMR) with spectral editing. Regardless of synthesis conditions, hydrochars consist mostly of oxygensubstituted arene rings (including diphenols) and furans connected by alkyl linkers rich in ketones. Cross-linking nonprotonated and methyne (C-H) alkyl carbons have been identified through spectrally edited 2D NMR. Alkenes and 'quaternary' C-O are observed only at low synthesis temperature, while some clusters of fused arene rings are generated at high temperature. Hydrochar composition is nearly independent of reaction time in the range from 1 to 5 h. Equilibration of 13C magnetization within 1 s shows that the materials are homogeneous on the 5-nm scale, refuting core-shell models of hydrochar microspheres. While furan C-O carbons bonded to alkyl groups or ketones show distinctive cross peaks in 2D NMR, phenolic C-OH is observed unambiguously by hydroxyl-proton selection. While methylene-linked furan rings are fairly common, the signal previously assigned to furan C alpha-C alpha linkages is shown to arise from abundant, stable catecholic ortho-diphenols, whose HO-C=C-OH structure is proved by 2D13C-13C NMR after hydroxyl-proton selection. Quantitative 13C NMR spectra of low- and hightemperature hydrochars have been matched by chemical-shift simulations for representative structural models. Mixed phenol and furan rings connected by ketones and alkyl linkers provide good fits of the experimental spectra, while literature models dominated by large clusters of fused rings and with few phenols or alkyllinked ketones do not.

During the process of subcritical water hydrolysis of lignocellulosic compounds, such as brewers' spent grains, some toxic compounds can be formed due to the breakdown of the cellulose structure. These compounds are known as furanic aldehydes, and they are inhibitors of microorganisms that are used in different processes, such as fermentation and anaerobic digestion, among others. The detoxification and conditioning of these hydrolysates are required before their use in other processes that require low concentrations of toxins. One of the mechanisms that can be used to diminish those compounds is the adsorption process. For this reason, this study evaluated at-line and in-line purification strategies using different adsorbents in fixed-bed columns to remove furanic aldehydes from a model feed simulating hydrolysate from subcritical water hydrolysis. The model feed consisted of simple sugars and furanic aldehydes at appropriate concentrations. The selected adsorbents were silica-C18, hydrocarbons, and activated carbon. The results indicate that, among the adsorbents that were studied, activated carbon was the most efficient for the removal of both 5-hydroxymethylfurfural and furfural. The activated carbon removed >99% of furanic aldehydes in the experiments conducted in the at-line system. Sequential subcritical water hydrolysis followed by either at-line or in-line purification is promising for integrated inexpensive and efficient hydrolysate purification.

Bio-based activated carbons with very high specific surface area of >3.000 m² g−1 (based on CO2 adsorption isotherms) and a high proportion of micropores (87% of total SSA) are produced by corncobs via pyrolysis and chemical activation with KOH. The activated carbon is further doped with different proportions of the highly pseudocapacitive transition metal oxide RuO2 to obtain enhanced electrochemical properties and tune the materials for the application in electrochemical double-layer capacitors (EDLC) (supercapacitors). The activated carbon and composites are extensively studied regarding their physico-chemical and electrochemical properties. The results show that the composite containing 40 wt.% RuO2 has an electric conductivity of 408 S m−1 and a specific capacitance of 360 Fg−1. SEM-EDX, XPS, and XRD analysis confirm the homogenous distribution of partly crystalline RuO2 particles on the carbon surface, which leads to a biobased composite material with enhanced electrochemical properties.

Regression ensembles consisting of a collection of base regression models are often used to improve the estimation/prediction performance of a single regression model. It has been shown that the individual accuracy of the base models and the ensemble diversity are the two key factors affecting the performance of an ensemble. In this paper, we derive a theory for regression ensembles that illustrates the subtle trade-off between individual accuracy and ensemble diversity from the perspective of statistical correlations. Then, inspired by our derived theory, we further propose a novel loss function and a training algorithm for deep learning regression ensembles. We then demonstrate the advantage of our training approach over standard regression ensemble methods including random forest and gradient boosting regressors with both benchmark regression problems and chemical sensor problems involving analysis of Raman spectroscopy. Our key contribution is that our loss function and training algorithm is able to manage diversity explicitly in an ensemble, rather than merely allowing diversity to occur by happenstance.

A model hydrochar was synthesized from glucose at 180 °C and its Cu(II) sorption capacity was studied experimentally and computationally as an example of molecular-level adsorbent design. The sorption capacity of the glucose hydrochar was less than detection limits (3 mg g−1) and increased significantly with simple alkali treatments with hydroxide and carbonate salts of K and Na. Sorption capacity depended on the salt used for alkali treatment, with hydroxides leading to greater improvement than carbonates and K+ more than Na+. Subsequent zeta potential and infrared spectroscopy analysis implicated the importance of electrostatic interactions in Cu(II) sorption to the hydrochar surface. Computational modeling using Density Functional Theory (DFT) rationalized the binding as electrostatic interactions with carboxylate groups; similarly, DFT calculations were consistent with the finding that K+ was more effective than Na+ at activating the hydrochar. Based on this finding, custom-synthesized hydrochars were synthesized from glucose-acrylic acid and glucose-vinyl sulfonic acid precursors, with subsequent improvements in Cu(II) adsorption capacity. The performance of these hydrochars was compared with ion exchange resins, with the finding that Cu(II)-binding site stoichiometry is superior in the hydrochars compared with the resins, offering potential for future improvements in hydrochar design.

Hydrochar was studied using Raman spectroscopy to understand thermal damage potentially introduced during analysis. Analysis of glucose hydrochar at power densities less than 5 GW m(-2) indicated a furanrich material, while analysis at power densities greater than 50 GW m(-2) indicated an arene-rich structure. Thermal tests using conventional heating revealed furan-to-arene transition in the tempera ture range between 300 and 500 degrees C, suggesting that Raman analysis at power densities greater than 5 GW m-2 converts furans to arenes. A heat transfer model predicted temperatures consistent with the transition identified experimentally and indicated pelletizing the hydrochar as an approach for acquisition of artifact-free Raman spectra without sacrificing signal-to-noise. Subsequent experimental analysis of hydrochar-KBr pellets confirmed the potential effectiveness of the approach. The artifact-free Raman spectrum of glucose hydrochar was re-analyzed and found to be much more consistent with structures inferred from solid state nuclear magnetic resonance (NMR) than previously thought. Hydrochar is a thermally sensitive material and should be studied with great care when subjected to temperatures greater than 300 degrees C or (equivalently) laser powers greater than 5 GW m(-2). The approaches described here should be used for future work in order to get a proper description of char structure. (C) 2020 Elsevier Ltd. All rights reserved.

Coking behavior of ZSM-5 was studied in the absence and presence of supercritical water (SCW) using dodecane cracking as a model reaction. SCW suppresses coke formation and inhibits conversion of coke precursors into polycyclic aromatic hydrocarbons. Interestingly, the thermal and structural characteristics of coke formed in the presence and absence of SCW appear to follow similar reaction trajectories, with the primary difference being rate inhibition in the presence of SCW. Further experiments indicated that SCW must suppress coke by an indirect role of catalyst modification and by a direct role involving chemical or physical participation in the coke formation mechanism. These results provide new insight into the role of SCW on coke formation and motivate future work to identify strategies for stabilizing acid sites in the presence of liquid water.

The hydrothermal pretreatment route is gaining research interest as a potentially green method for deconstructing lignocellusic biomass. Based on the relevant literature, the conversion of biomass into platform chemicals or energy carriers through hydrothermal processes has been found to be advantageous by reason of enhanced process performance, while being environmentally friendly and technologically innovative. In this review, an assessment has been made of recent research findings and reservations in regard to the synthesis of subcritical and supercritical hydrolysates and the production of platform chemicals namely ethanol, butanol, furfural, hydroxymethylfurfural, lactic acid, levulinic acid and its derivatives, succinic acid, sorbitol, and xylitol. This review also proposes a number of future research-oriented directions to harness the findings of primary research-oriented efforts for developing technically and economically feasible large-scale systems.

Fast pyrolysis bio-oils from biomass can potentially integrate with petroleum refinery infrastructure for production of renewable fuels and chemicals. Besides hydro-deoxygenation, few feasible options exist for entry points. When considering advanced pyrolysis techniques such as catalytic and/or tail-gas reactive pyrolysis (TGRP), distillation for using both light and heavy ends becomes possible. Our goal was to demonstrate and optimize continuous production of liquid organic distillates and residual solids coke, both in appreciable yields for downstream conversion into renewable products. We fabricated a flash drum for continuous one-step distillations of four oils of varying oxygen content (ranging from 5 to 32 wt %). While a mesh demisting screen enhanced separation, removal of the screen ultimately improved overall yields. The flash drum proceeded to distill lower-oxygen oils (similar to 10 wt %) with 80 wt % time-on-stream yields over several hours; steady state was reached within 30-40 min. Bio-oils with moderate oxygen levels (20 wt %) took a noticeably longer time to attain steady state and gave 60 wt % yield. Under distillation conditions, oils from conventional pyrolysis (32 wt %) underwent condensation repolymerization due to reactive instabilities and produced only 6 wt % organic liquid yield. Solid coke residues were collected and converted into calcined coke, with Raman analysis indicating that catalytic and/or TGRP oil residues had higher molecular weight polyaromatics than those from traditional oil.

This study presents the effect of lignocellulosic compounds and monolignols on the yield, nanostructure and reactivity of soot generated at 1250 degrees C in a drop tube furnace. The structure of soot was characterized by electron microscopy techniques, Raman spectroscopy and electron spin resonance spectroscopy. The CO2 reactivity of soot was investigated by thermogravimetric analysis. Soot from cellulose was more reactive than soot produced from extractives, lignin and monolignols. Soot reactivity was correlated with the separation distances between adjacent graphene layers, as measured using transmission electron microscopy. Particle size, free radical concentration, differences in a degree of curvature and multi-core structures influenced the soot reactivity less than the interlayer separation distances. Soot yield was correlated with the lignin content of the feedstock. The selection of the extraction solvent had a strong influence on the soot reactivity. The Soxhlet extraction of softwood and wheat straw lignin soot using methanol decreased the soot reactivity, whereas acetone extraction had only a modest effect.

This review summarizes the recent essential aspects of subcritical and supercritical water technology applied tothe extraction, hydrolysis, carbonization, and gasification processes. These are clean and fast technologies which do not need pretreatment, require less reaction time, generate less corrosion and residues, do not usetoxic solvents, and reduce the synthesis of degradation byproducts. The equipment design, process parameters, and types of biomass used for subcritical and supercritical water process are presented. The benefits of catalysis to improve process efficiency are addressed. Bioactive compounds, reducing sugars, hydrogen, biodiesel, and hydrothermal char are the final products of subcritical and supercritical water processes. The present review also revisits advances of the research trends in the development of subcriticaland supercritical water process technologies.

Density Functional Theory (DFT) and experimental measurements were used to develop a systematic method for interpreting the Raman spectra of hydrothermal char (hydrochar). Average band locations, relative intensities, and their trends relative to structural features were determined for the G, D, and Kekulé bands. When combined with several other less prominent vibrational modes, including vibrations associated with aromatic, ether, alkyl, and carbonyl bonds, the calculated average locations reproduced all major features of hydrochar Raman spectra. Two model structures were found that could reproduce the main features of the hydrochar Raman spectrum and its elemental analysis: 1) a structure consisting of arene domains comprised of 6–8 rings connected via aliphatic chains or 2) a furan/arene structure consisting primarily of single furans and 2 or 3ring arenes. NMR confirmed that the furan/arene ratio of glucose hydrochar is approximately 1:1, consistent with the furan/arene structure supported by Raman spectra. This work establishes an interpretation method for hydrochar Raman spectra and reconciles the main features of hydrochar structures determined using Raman spectroscopy with those based on other methods.

As the electric power grid evolves to a smart grid, the introduction of real-time price structures will provide consumers with the opportunity to obtain low cost (possibly negative cost) electrical energy. This work will propose a novel design for a utility plant within a chemical processing facility that will enable exploitation of the diurnal nature of expected electricity prices. Specifically, we will investigate an oil heating, gas fired furnace that has been augmented with an electric heater. A second configuration, that augments the electric heater with an energy storage unit, will also be investigated. Results indicate that substantial savings in energy cost can be achieved, but will likely be undercut by the capital costs associated with the electric heater.

The effect of lignocellulosic compounds and monolignols on the soot nanostructure and CO2 reactivity